Abstract
Privacy and security in the parameter transmission process of federatedlearning are currently among the most prominent concerns. However, there aretwo thorny problems caused by unprotected communication methods:"parameter-leakage" and "inefficient-communication". This article proposesBlockchain-based Federated Learning (FBChain) model for federated learningparameter communication to overcome the above two problems. First, we utilizethe immutability of blockchain to store the global model and hash value oflocal model parameters in case of tampering during the communication process,protect data privacy by encrypting parameters, and verify data consistency bycomparing the hash values of local parameters, thus addressing the"parameter-leakage" problem. Second, the Proof of Weighted Link Speed (PoWLS)consensus algorithm comprehensively selects nodes with the higher weighted linkspeed to aggregate global model and package blocks, thereby solving the"inefficient-communication" problem. Experimental results demonstrate theeffectiveness of our proposed FBChain model and its ability to improve modelcommunication efficiency in federated learning.